Front page | perl.perl5.porters |
Postings from July 2001
On Wed, 4 Jul 2001, Michael G Schwern wrote:
> A question. A good chunk of MM_VMS.pm is just replacing this style:
>
> $(PERL) -I$(INST_LIB) -e 'whatever'
>
> with this style
>
> $(PERL) "-I$(INST_LIB)" -e 'whatever'
>
> Since I believe the later is perfectly acceptable on Unix, could we
> just fix up MM_Unix to do that and chuck most of MM_VMS?
You might want to try that out (moving some of the double quotation stuff
to MM_Unix.pm). Since we are now doing things like:
`$^X "-I../lib" -e 'whatever'`;
within the perl regression tests there might not be too bad an impact on
other platforms (the above will work on most Unices, NT, & VMS; the
'../lib' would need changing to '::lib:' for Mac OS).
There is considerably more to MM_VMS than just the double quotations. In
fact I would assert that much of MM_VMS.pm is:
0) getting around DCL's 255 character command line limit (that affects
command lines run from MMS or
MMK make utilities)
1) changing 'alist.o of_object.o files.o' to
'alist$(OBJ), of_object$(OBJ), files$(OBJ)' (note the use of the commas in
the latter)
2) creation of linker options files where steps 0 and 1 have
still not granted short enough link lines.
There is more too in that 2>&1 can work from perl (since it is handled in
vms/vms.c but the shell does not do that since it has:
"\$ set message/nofacil/nosever/noident/notext\n";
which pretty unique to the default command language interpreter on VMS.
Hence it may prove difficult or impossible to get rid of MM_VMS.pm,
certainly without littering MM_Unix.pm with oodles of C<if ($^O eq 'VMS')>
which kind of defeats the whole code re-use idea of MakeMaker.
Peter Prymmer
Thread Previous
|
Thread Next